User Guide

The user guide covers each stage of the ALSEBO pipeline in depth. Read the pages in order for a full conceptual walkthrough, or jump directly to the component you need.

• Overview
  • What is ALSEBO?
  • The ALSEBO pipeline
  • Modules at a glance
  • Choosing a featurization method
  • References
• VAE Latent Landscape
  • Conceptual background
  • Architecture
  • Training
  • Generating the sequence space
  • Latent landscape as a functional map
  • References
• Sequence Space
  • Featurization methods
  • Output format
• Training Space
  • Why diversity matters
  • Sampling strategy
  • Recording experimental results
  • The obj_config dictionary
• Bayesian Optimisation
  • The surrogate model — Gaussian Process Regression
  • The acquisition function — UCB
  • Tuning beta
  • Appending results and closing the loop
  • Round-by-round improvement